AI-Enabled Cyberattacks Accelerating — Anthropic Policy Stance Under Fire
Three likely AI-enabled cyberattacks in 10 days undercut prior dismissals of near-term AI security threats, while critics argue Anthropic cares more about avoiding government shutdown authority than about AI safety.
What we know
an account on X (amplified by both two accounts on X) notes that a recent argument downplaying AI cyberattack risks 'has aged incredibly badly' given three significant likely AI-enabled attacks in ten days (tweet). A separate KelseyTuoc thread argues Anthropic — while acknowledging they are doing something 'extremely dangerous' — is primarily motivated in its policy positions by fear of giving the executive branch flexible authority to shut it down (tweet). Separately, another account on X highlights a video showing how malware using a legitimate filename but running under a random user account reveals a live foreign attack trace (tweet).
By the Global Eye 24 desk, with AI assistance. Updated as the story develops. Spot an error? Write to corrections@globaleye24.com.